基于光谱角匹配算法的水体信息提取研究

Water Extraction Based on Spectral Angle Matching Algorithm

针对水体信息提取时,传统算法不能较好地解决山体阴影被误提为水体的问题,引入多用于高光谱数据处理的光谱角匹配算法,以Land-sat8 OLI(band1-band7)为数据源,以黄河小浪底水库周边区域为实验区,开展水体信息提取研究,结果显示:水体指数法和波段关系提取的结果中有大量的山体阴影信息,提取精度不超过30%,而光谱角匹配法提取结果受阴影的影响较小,提取精度在99%以上。

Traditional water extraction method can＇ t distinguish mountain shadow and water body, to solve this question, the spectral angle matching algorithm which is used to process the hyper spectral data usually was introduced, using this algorithm, the study on water information extraction was carried out in the surrounding area of the Yellow River Xiaolangdi reservoir. The study result of water index and the relationship between multi - band spectral methods contained a lot of mountain shadow information, the extraction accu- racy was less than 30%, however, the spectral angle matching algorithm had achieved better results, and the extraction accuracy was out of 99%.